Feedback loops social media platforms use to maximize engagement often amplify outrage and division as a side effect of the loop’s design.
Engagement as the Optimized Variable
Recommendation algorithms are tuned to maximize time on platform, and the fastest way to hold attention is content that triggers strong emotion. Feedback loops social media systems create reward exactly this, regardless of the content’s accuracy or civility.
The Reinforcing Loop of Outrage
Outraged reactions generate more comments and shares, which the algorithm reads as high engagement, which pushes the content to more people, who then react with more outrage. This loop compounds quickly once it starts.
Filter Bubbles as a Structural Outcome
Personalization loops narrow the range of content a user sees over time, not through explicit censorship but through the accumulating weight of past engagement data steadily reshaping the feed.
Interventions That Target the Loop Itself
Friction like reshare warnings or slower notification cadence works by weakening the reinforcing loop directly, rather than trying to moderate every individual piece of content after the fact.
